Empirical Study of Different Classifiers for Sentiment Analysis

The usage of textual or unstructured data has increased rapidly in the present day scenario. Nowadays websites, social networking as well as many organizations use this sort of data. A major problem occurs when we try to determine the sentiment or the class of these data i.e. whether the data is good or bad. Analyzing the sentiment of a text, document or an article is a challenging task in the world. Several methods were implemented for sentiment analysis throughout the years, but still more improvement and perfection is needed. In this paper, some sentiment based datasets were taken along with a dataset created from reviews collected from Flipkart, a popular online shopping website was also used and a sentiment based function is implemented and finally some classifiers like Naive Bayes, Support Vector Machines (SVM), decision tree and k-Nearest Neighbors (k-NN) were used to predict the accuracy of determining the sentiment type or the class. The objective of this paper is to analyze the accuracy and performance of different classifiers.